UNGULATE RESPONSES TO MULTI-USE PATHWAY CONSTRUCTION AND USE
IN GRAND TETON NATIONAL PARK, WYOMING
FINAL REPORT
Prepared for the National Park Service, Grand Teton National Park
September 15, 2011
Amanda R. Hardy, PhD Candidate
&
Kevin R. Crooks, Associate Professor
Graduate Degree Program in Ecology
Department of Fish, Wildlife and Conservation Biology
Colorado State University
Fort Collins, Colorado 80523
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
1
ABSTRACT
In 2006 Grand Teton National Park (GTNP) adopted a transportation plan which includes the
construction of paved, multi-use pathways along existing park roads to accommodate non-
motorized travelers, raising concerns about potential impacts of the pathway on wildlife and
wildlife viewing experiences. We initiated a repeated measures Before-After-Control-Impact
(BACI) study to assess the potential effects of pathway construction and use on the distribution
and behavior of ungulates (elk, pronghorn antelope, moose, mule deer) and on wildlife viewing
opportunities. Data collection occurred before pathway construction (2007), during construction
(2008) and for two years after the pathway was open to public use (2009 and 2010) in a 6.9 km
control region (without pathway) and a 12.5 km treatment region (with pathway) along Teton
Park Road. If ungulates responded to pathway construction and use, we predicted that, in the
treatment as compared to the control: 1) standardized counts of ungulates viewed from the road
would decline; 2) ungulates would be detected farther from the road; and 3) the probability of
ungulates responding behaviorally would be higher. Further, if ungulates avoided pathway
activities, we expected wildlife viewing activities to decrease in the treatment compared to the
control after pathway installation. During 20 months of data collection over the four years of the
study, we conducted 421 road surveys, collecting observations during 5,126 sampling miles
driven, recording 23,424 human activities and 2,319 ungulate groups comprised of 14,769
observations of individual ungulates; additionally, we conducted 670 focal samples recording
151 hours of ungulate behaviors. Park visitors stopped and were seen afoot and on bicycles in
the treatment more often than in the control. When the pathway opened in 2009, bicyclists
switched from using the road to using the path, and there was an overall increase in bicycling and
pedestrian activities on the pathway in 2009 and 2010. More people were observed viewing
wildlife in the treatment compared to the control throughout the study. Notably, we saw a three-
fold increase in off-road human activity in the treatment after the pathway was constructed, an
activity with a greater potential to disturb wildlife than activities that remain on the road and
pathway. We observed elk most frequently, followed by pronghorn, while we observed moose
and mule deer relatively infrequently. Moose and mule deer were observed in the treatment
more frequently than the control, and did not seem to be displaced by pathway activities,
although observations were limited. During early season (June-July 15), the number of elk
viewed in the treatment was stable or increased slightly, while elk viewed in the control
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
2
decreased between 2008 to 2010. In mid season (July 16-August 31), when park visitation and
pathway use peaked annually, elk appeared to less be behaviorally responsive in the treatment
compared to the control after pathway installation. These results are contrary to predictions that
elk would avoid pathway activities and suggest that elk might have developed tolerance to
human activities in the treatment. During mid season, pronghorn were seen, on average, about
164 meters farther from the road in the treatment in 2010 compared to 2007, but shifted closer to
the road in the control area over the same period. This suggests that pronghorn did to a limited
degree avoid habitats proximate to the road where the pathway was installed, supporting one of
our predictions indicative of a pathway effect. Overall, while pathway construction and use
resulted in direct habitat loss and widened and diversified the human footprint, our results did not
consistently demonstrate alterations in ungulate distribution and behavior or wildlife viewing
opportunities, with the exception of a relatively modest shift in pronghorn distribution, and an
apparent decrease in the behavioral responsiveness of elk, during periods of peak visitation and
pathway use.
INTRODUCTION
Transportation systems in national parks allow millions of people to see and experience an array
of natural and cultural resources that park managers aim to protect in perpetuity. Roads and
automobile traffic in parks and elsewhere impose negative impacts on natural terrestrial and
aquatic systems, destroying and altering habitats, spreading non-native biota (Trombulak and
Frissell 2000), modifying and creating barriers to animal movements (Riley et al. 2006), masking
and interfering with animal auditory communications and sensing (Barber et al. 2010), and
killing organisms that attempt to cross roads but are struck by vehicles (Trombulak and Frissell
2000, Evink 2002, Forman 2003). Road impacts on wildlife are considered a concern for park
managers (Ament et al. 2008). To reduce motor vehicle impacts on park resources, diversify
recreational experiences, and increase safety for non-motorized travelers, there is a growing
interest in alternative transportation options such as recreational pathways to complement park
roads.
While the inclusion of pathway systems accommodates non-motorized travel for park visitors,
with the potential to reduce traffic congestion and decrease bicycle-automobile conflicts, such
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
3
infrastructure widens travel corridors and extends the influence of humans into habitats that may
be important to wildlife. Human disturbance has been shown to affect different ecological
characteristics of wildlife, such as home range size and habitat use, foraging behavior,
reproductive success, body condition, disease susceptibility, sex ratio, daily activity period,
social development, mating systems and social structure (see review in Bejder et al. 2009). In
parks where wildlife are seen near roads, recreational activities on pathways may prompt local
wildlife to change activity patterns, avoid habitats near recreational pathways, and increase
vigilance and energy expenditures (Knight and Cole 1995, Borkowski et al. 2006, George and
Crooks 2006). Adding a recreational pathway along a well-traveled road known for wildlife
viewing opportunities is unprecedented in a US national park (National Park Service 2006), thus
it is uncertain how this new infrastructure might affect wildlife and wildlife viewing
opportunities and experiences.
In 2006 Grand Teton National Park (GTNP) adopted a Transportation Plan that calls for a paved,
multi-use pathway system to accommodate bicyclists, pedestrians and other non-motorized travel
modes (National Park Service 2006). Under the plan, approximately 30.3 km (18.8 miles) of
existing roadways could be reconstructed and widened to include multi-use pathways, while
another 36.0 km (22.5 miles) of paved pathways could be installed 15.2 to 45.7 m (50 to 150
feet) from existing roads that traverse open habitats occupied by large ungulates (pronghorn
antelope [Antilocapra americana], elk [Cervus elaphus], moose [Alces alces], and mule deer
[Odocoileus hemoinus]) which may be seen from park roads during summer and fall months.
This pathway would widen and diversify the “human footprint”, extending visitor activities
further into wildlife habitats and raising concern about how this new infrastructure may affect
behaviors, movements and interactions of animals and park visitors coinciding in these park
travel corridors. Park management contracted several independent studies to quantify and assess
effects of the first phase of pathway construction (12.5 km, 7.7 miles, constructed in 2008) and
use on birds, bears and ungulates, including this study, which focuses on ungulates, park visitors
and their interactions in the pathway corridor.
The goal of this field research was to assess if and how pathway construction and use affects
distribution and behavior of pronghorn, elk, moose and mule deer, as well as visitors’ wildlife
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
4
viewing opportunities and interactions with wildlife in GTNP. To accomplish this goal, we
implemented a repeated measure environmental impact Before-After-Control-Impact (BACI)
study design (Stewart-Oaten et al. 1986, Green 1993, Stewart-Oaten and Bence 2001, Smith
2002, Roedenbeck et al. 2007). Data collection occurred before pathway construction (2007),
during construction (2008) and for two years after the pathway was open to public use (2009 and
2010) in a control region (without pathway) and a treatment region (where the pathway was
constructed). We sampled human activities along with the number of ungulates observed in the
control and treatment areas, how close ungulate groups were to the transportation corridor and
the behaviors of ungulates to understand if they were affected by pathway activities. If ungulates
avoided pathway construction and use, we predicted that during and after pathway construction,
compared to the control: 1) standardized counts of ungulates observed in the treatment would
decline, 2) ungulates would be detected farther from the transportation corridor in the treatment,
and 3) ungulates would exhibit increased behavioral responsiveness in the treatment area. We
also expected to see changes in the types and frequencies of human activities occurring in the
treatment area with the installation of the pathway. If wildlife were negatively affected by these
changes and avoided habitats near the road and pathway, we would expect to see fewer people
watching wildlife in the treatment area after the pathway was constructed and opened.
STUDY AREA
This study was conducted in Grand Teton National Park (GTNP) in northwestern Wyoming,
USA (43-50'00'' N, 110-42'03'' W). The study area included habitats visible from the 19.4 km
(12 miles) of the Teton Park Road (TPR) between Moose and Spalding Bay Road, at the
southern and northern extents of the study area, respectively (Figure 1). This section of the TPR
traversed the valley floor between the Teton Mountains to the west and the Snake River to the
east; study area landscapes visible from the TPR spanned elevations from 2133 meters (6998
feet) at the northern end of the study area to 1962 meters (6437 feet) at the southern end of the
study area. The landscape visible from the TPR was characterized by flat, open, glacial outwash
valley plains with vegetation dominated by dry sagebrush (Artemesia spp.) shrublands with
patches of lodgepole pine (Pinus contorta), Douglas fir (Pseudotsuga menzesii), Engelman
spruce (Picea engelmannii) and smaller stands of aspen (Populus tremuloides) woodlands
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
5
providing cover for wildlife. Willow (Salix spp.) riparian zones and cottonwoods (Populus
angustifolia) lined the streams, ponds and remnant irrigation water features in the study area.
Figure 1. The study region included habitats visible from 19.4 km (12 miles) of the existing Teton Park Road
(TPR) between Moose and Spalding Bay Road (marked with stars, inset a). Viewshed analysis estimated area
visible from 42 established scan points along the TPR in both areas (control scan points =17; treatment scan
points = 25; inset b). In 2008, construction of 12.5 kilometers (7.7 miles) of paved pathway next to the road
created a “treatment area” (solid ellipses) while the northern 6.9 km (4.3 miles) of the TPR in the study
region had no pathway and served as a “control area” (dashed ellipses).
The study area provided summer and fall habitat for resident and migrating ungulates, including
elk, pronghorn antelope, mule deer, and moose. Grizzly and black bear (Ursus arctos horribilus
and Ursus americanus), grey wolves (Canis lupus), mountain lions (Felis concolor), coyotes
a)
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
6
(Canis latrans) and red fox (Vulpes vulpes) have been known to use study area habitats as well.
Additionally, thousands of park visitors pass through the study region on the TPR each year,
with a peak in visitation typically occurring in July. While most people travel through the study
area on the TPR in an automobile, a small portion of visitors opted for non-motorized travel
modes and recreation (e.g., bicycling, walking).
In 2008, 12.5 kilometers (7.7 miles) of pathway was constructed along the TPR in the southern
end of the study area to accommodate non-motorized travel between South Jenny Lake and
Moose. The installation and public use of the 3.05 meter (10 foot) wide paved route along the
existing road created a “treatment” region within the larger study area (Figure 1). The northern
end of the study area, along 6.9 km (4.3 miles) of the TPR between South Jenny Lake Junction
and Spalding Bay Road, provided a “control” site where no pathways were constructed during
this study. Although the control and treatment areas similarly shared the same habitats and types
of topographical features, the two adjacent areas were not identical in size or configuration.
These distinctions were considered in our analyses and interpretation of results.
METHODS
We applied a repeated measures Before-After-Control-Impact (BACI) approach to assess the
potential effects of pathway construction and use on ungulates (Stewart-Oaten et al. 1986, Green
1993, Stewart-Oaten and Bence 2001, Smith 2002, Roedenbeck et al. 2007). We conducted
systematic road surveys and focal sampling methods to repeatedly measure variables of interest
in the treatment and control area before (2007), during (2008) and after (2009-2010) the pathway
installation and use in the treatment region. Our assessment focused on four response variables:
1) numbers of ungulates viewed (standardized by viewable area) per survey; 2) distance of
ungulate groups to TPR; 3) probability of ungulates behaviorally responding during a group
behavior scan; and 4) probability of ungulates behaviorally responding during focal animal
sampling. We also sampled human activities occurring in the study area during road surveys as
an index of potential change in stimuli which might be related to ungulate responses. If trends in
the response variables were similar in the control and treatment area over the years (before,
during and after pathway introduction), it would imply that the pathway, introduced into the
treatment area in 2008, did not influence the response variables. Alternatively, if trends in
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
7
response variables were different in the treatment and control area after pathway installation, this
would then suggest that ungulates may have responded to pathway activities.
Field methods
Road survey methods involved two observers systematically traversing the 19.4 km (12 mile)
TPR study area, from Moose to Spalding Bay Road (or vice versa; survey direction was
alternated), to sample human and ungulate activities which could be observed from the road.
Surveys were conducted typically 1-2 times a day from June through October, with start times
staggered 12-14 hours (e.g., Day 1: 6am & 6pm; Day 2: 8am, 8pm, etc.). Surveys typically
spanned 2-4 hours depending on ungulate activity in the corridor. Some surveys spanned both
light and dark periods (as defined by sunset/sunrise tables); however, most surveys were initiated
and completed during daylight conditions.
Observers traveled in a truck at ~48 km/h (~30 mph) through the study area, stopping at 42 “scan
points” established approximately ~160-804 meters (~0.1-0.5 miles) apart in select locations
where views of the landscape were maximized (Figure 1). Observers also stopped
opportunistically to record data (described below) if ungulates were observed between scan
points. At each scan point, immediately upon stopping, and prior to our presence and activities
potentially influencing visitor behavior, observers sampled human activities occurring within
200 m of the scan point. This included recording the number of autos stopped on the side of the
road (i.e., gravel shoulder) and in paved pullouts, the number of pedestrians (i.e., people afoot)
and bicyclists on the side of the road and in paved pullouts, and the number of pedestrians and
bicyclists on the pathway in the treatment in 2009 and 2010. Non-motorized activities were
further distinguished by movement (passing versus stopped) when first observed. In 2008, 2009,
and 2010, we recorded whether people were viewing wildlife and all human activities off-road
and off-pathway. If ungulates were visible when the observers first stopped at a scan point,
observers would note the majority of the group’s behavior and approximate location when first
seen, in case the group disappeared while human activities were being recorded.
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
8
After recording human activities, observers searched the landscape for ungulates using
binoculars. An ungulate group was defined as ≥1 ungulate(s), with nearest neighbor distances
between ungulates <100 meters (328 ft) within groups and >100 meters between groups. When
ungulate groups were observed at any point during road surveys (i.e., at and between scan
points), observers recorded the time of observation, species, location, and the total number of
animals in the group. Location of the group was determined based on the observers’ location,
distance between observers and the ungulate group, and a compass azimuth to the center of the
group. Distance from observers to the group was determined using a laser range finder; if the
group was beyond the effective range of the laser range finder, we used the azimuth to the group
and a topographic map to estimate a distance from observer to group. If the group was observed
within ~500 meters of the TPR, one observer scanned and recorded each animal’s behavior in the
group. These behaviors were categorized in a manner similar to definitions by Childress and
Lung (2003) and Borkowski et al. (2006), as follows: bedded; feeding; grooming (i.e., licking or
scratching oneself or another); scanning (i.e., standing with head at or above shoulder level, but
not apparently alarmed); traveling (i.e., walking); vigilant (i.e., displaying alarm or acute
attention toward some stimulus); flight (i.e., running away from some stimulus); defensive (i.e.,
kicking, biting or charging towards another animal); and mating (i.e., rut behaviors observed in
the fall months such as grouping and pursuing cows or sparring between bulls).
In addition to recording scan behaviors of each ungulate in a group, we conducted focal animal
samples, collecting continuous behavioral observations of individual animals both during road
surveys and opportunistically (e.g., prior to and after surveys); we avoided collecting >1 focal
sample from the same group in a day, as possible. We adapted focal animal sampling methods
from Childress and Lung (2003) When an ungulate group was sighted within 500 meters of the
road, we collected the same attributing data as described in road survey methods. One observer
randomly selected an individual from the group and continuously recorded behaviors categorized
as described above for approximately 15 minutes during road surveys. If the focal animal
sample occurred opportunistically before or after a road survey, observers continued focal
sampling as long as possible or until the focal animal moved out of sight.
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
9
Data Analysis
We calculated annual means and standard errors (SEs) of human activities observed within 200
meters of scan points in the treatment and control by summing each category of activities seen in
each area per survey and dividing that value by 25 or 17 scan points in the treatment or control
area, respectively. Wildlife viewing activity was reported as present or absent at each scan point;
annual summaries of this activity are reported as the mean percent of scan points where wildlife
viewing was occurring per survey, by area (i.e., treatment and control).
Individual road surveys conducted during daylight and crepuscular periods served as the
experimental unit to analyze numbers of ungulates viewed. To standardize number of ungulates
viewed in the treatment and control given that these areas were not identical in size or
configuration, we divided the count of ungulates viewed, per survey, by hectares (ha) of habitat
that could be seen from the road in each area. The area of habitat visible from the road was
estimated using a viewshed analysis in a Geographic Information System (GIS; ArcMap v9.3,
ESRI, Redlands, CA); see inset b in Figure 1 for representation of viewshed output. The
estimation of hectares of visible terrain accounted for differences in the control and treatment
regions, and for different heights (hence, different detectability) of large (moose, elk) and small
(mule deer, pronghorn) ungulates. In the control area, elk and moose counts for each survey
were divided by 986 ha (3.81 square miles) of visible terrain, whereas deer and pronghorn counts
were divided by 932 ha (3.60 square miles). In the treatment area, elk and moose counts per
survey were divided by 2,261 ha (8.73 square miles) and pronghorn and deer counts were
divided by 2,111 ha (8.15 square miles). We refer to this standardized measure (views per
viewable area) as “animals viewed”, and graphically display mean numbers and standard errors
of animals viewed per survey for effects plots.
We used observations of ungulate groups during daylight and crepuscular conditions as the
experimental unit to analyze distances of ungulates to the TPR. We used a GIS to determine
perpendicular distance (meters) of each group to the TPR based on the group’s location. Means
and standard errors of perpendicular distances were presented in effects plots. Distance values to
the TPR exceeding 1000m were difficult to accurately estimate. Thus, for statistical analyses,
values exceeding 1000m were reclassified to equal 1000m to reduce variability in the data while
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
10
retaining observations of groups that could be seen at extremely far distances from some points
on the TPR.
For group scan and focal animal sampling behavioral analyses, we created a binomial response
variable, pooling flight, vigilance, traveling and defensive behaviors as “responsive” behaviors,
and bedded, feeding, grooming, scanning and mating behaviors as “non-responsive” behaviors
(Goldstein 2005). For group scan data, this binomial response variable estimated the probability
that an individual within a herd was responsive. For focal animal data, this binomial response
variable measured the probability that the focal animal was responding each second during the
focal sampling period. For scan samples, we created effects plots of the annual mean (and SE)
proportion of animals responding per group (number responsive / total number animals in
group). Similarly, for focal animal samples, we plotted the mean (and SE) proportion of time an
individual was responding per focal sample (number of seconds responsive / total seconds in the
focal sample).
We modeled the effect of year (2007, 2008, 2009, 2010), area (treatment, control) and time (day,
crepuscular) on each response variable (ungulates viewed per survey, distance from road per
group, probability of an individual during group behavior scan, and probability of an individual
responding during focal sample). Time was defined as “day” if the sample occurred entirely
during daylight conditions ( 1 hour after dawn and 1 hour before dusk), or as “crepuscular” if
any portion of the survey or observation was recorded during crepuscular hours (≤1 hour after
dawn or ≤1 hour prior to dusk) as determined by regional sunrise and sunset tables (U.S. Naval
Observatory 2010).
We square root-transformed the two continuous response variables (animals viewed per survey,
distance from road per group) to meet the assumption of linearity and used a linear mixed model
to evaluate whether fixed effects and their interactions influenced numbers of ungulates viewed
and ungulate distances to the TPR, as follows:
Xijk = μ + αi + βj + γk + τl(i) + (αβ)ij+(αγ)ik +(βγ)jk + (αβγ)ijk + εijkl
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
11
where μ is the overall mean, αi is the fixed effect of year (i = 2007, 2008, 2009, 2010), βj is the
fixed effect of area (j = control(0), treatment(1)), γk is the fixed effect of time (k=day(0),
crepuscular(1)), (αβ)ij is the interaction between year and area, τk(i) is the random effect of survey
where k(i) = 1, …, n surveys conducted in year i, (αγ)ik is the interaction between year and time,
(βγ)jk is the interaction between area and time, (αβγ)ijk is the three-way interaction between the
three fixed effects, and εijkl is the error in the mean not explained by fixed effects. The random
effect of survey nested within year was included given that numbers of animals viewed and their
location relative to the road in the control and treatment areas during a given survey may be
influenced by environmental (e.g., weather, visibility, length and time of day) and biological
(e.g., inter- and intraspecific interactions, plant phenology) factors occurring during the survey.
For the binomial response variables, we used logistic regression with a logit link function to
assess whether year, area and time and their interactions predicted the probability of an
individual responding. For the group scan data, we again included a random effect for survey
nested within year; additionally, we included group nested within survey and year to account for
potential pseudoreplication issues of correlated behavioral responses within a group, given that
an individual’s response may be influenced by other group members’ behaviors. For the focal
animal data, we included a unique integer accounting for the random effect of the date nested
within year (as opposed to survey within year), given that focal samples were recorded both
during surveys and opportunistically before and after road surveys. This random effect accounted
for similar environmental factors (e.g., weather) and biological factors (e.g., presence of other
ungulates, predators, humans) shared on days when >1 focal sample was collected.
Each species was analyzed separately. When low sample sizes and high variability prevented
statistical modeling of a given response variable (e.g., in the moose and mule deer data), we
limited reported results to qualitative assessments of annual means (and standard errors), pooled
over seasons and time of day, and limited the interpretations of these results accordingly. We
conducted analyses of elk and pronghorn data over three seasons: early season (June to July 15,
when ungulates were calving/fawning), mid season (July 16 to August 31, when human
visitation and temperatures peaked), and late season (September 1 to October 15, when ungulates
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
12
were influenced by the rut, hunting pressure beyond the study area and oncoming winter weather
conditions).
Statistically significant (p<0.05) interactions indicated a trend observed in a given response
variable varied depending on differences in the two or three variables in the interaction. We
were particularly interested in significant year*area interactions as an indicator of different
responses observed in the treatment area as compared to the control area after pathway activities
were introduced. Thus, we ran additional post hoc Tukey-Kramer Honestly Significant
Difference (HSD) tests, comparing significant seasonal year*area interactions within each area
between years (2007-2008, 2007-2009, 2007-2010, 2008-2009, 2009-2010). We conducted
additional post hoc ANOVA analyses to further examine significant seasonal 3-way interactions
(year*area*time), excluding the main effect of time from the model and reanalyzing crepuscular
and day observations separately, to assess whether the year*area interaction was significant
during either time period. We used average values to further interpret significant contrasts
between years within an area and to interpret how divergent trends between the treatment and
control areas might be indicative of a pathway effect. Results reported in the main text focused
on statistically significant main effects and on the interactions incorporating year and area
(year*area and year*area*time). Tables with all parameter estimates and p-values for main and
post hoc analyses are referenced as well. We used PROC GLIMMIX in SAS (v9.2) for ANOVA
analyses and for calculating means and SEs for plotting, and created effects plots using R
(version 2.13.1) and Microsoft Office Excel 2007.
RESULTS
We began collecting field data on June 24th in 2007, and on June 5th, 9th, and 10th in the 2008,
2009 and 2010 field seasons, respectively. Our last day of data collection ranged between
October 13-16th over the four years of the study. During these four years, we conducted 421
road surveys (Table 1), systematically collecting observations of ungulate and human activities
over 5,126 miles driven across the study region. Most road surveys (n = 242 surveys, 57%) were
conducted completely during the day while a smaller proportion of road surveys (43%, n = 179
surveys) started or ended during crepuscular periods (Table 1).
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
13
Table 1. Road survey effort summarized by year and time of day. Surveys were conducted in Grand Teton
National Park between June-October of 2007-2010.
YEAR T
OT
AL #
surv
eys
Cre
puscula
r
surv
eys
Day s
urv
eys
2007 69 25 44
2008 95 45 50
2009 133 59 74
2010 124 50 74
2007-2010 421 179 242
Human Activities
We sampled human activities observed within 200 meters of 17 scan points in the control on
7,276 occasions and 25 scan points in the treatment on 10,700 occasions, producing 21,483
observations of human activities. A disproportionate number of human activities
(84%, n= 18,148 observations) were recorded during day surveys (57% of all surveys), while
only 16% (n=3,335 observations) of human activities were sampled during crepuscular surveys.
Over the four years of the study, there appeared to be a slight upward trend in the numbers of
automobiles observed stopped in the study area during surveys. Most stopped automobiles were
seen in pullouts in the treatment area, increasing from an average of 25 to an average of 30 seen
per road survey over the four years of the study (Figure 2). In comparison, we saw an average of
2-3 autos in pullouts in the control area per survey. Mean number of automobiles observed
stopped along the road per survey in the control was relatively low, from <1 to 2, compared to a
3-5 automobiles seen stopped along the road in the treatment during surveys.
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
14
Figure 2. Mean number (± standard error) of automobiles stopped within 200 meters of road survey scan
points in the control (n = 17 scan points) and treatment (n=25 scan points) areas during road surveys
conducted in Grand Teton National Park between June-October 2007-2011.
Most pedestrian activity was observed in treatment area pullouts (Figure 3). In the control, we
saw few pedestrians in pullouts (ca. 2) and afoot on the road (<1) on average, per survey. In the
treatment, the mean number of people observed afoot on the road decreased from 6 to 3 people
between 2007 and 2010. It was rare to see people off road in the control region, but we saw an
increase in people off road and off path in the treatment from about 1 to 3 people per survey
between 2008 and 2010 (Figure 3).
Figure 3. Mean number (± standard error) of people afoot within 200 meters of road survey scan points in
the control (n = 17 scan points) and treatment (n=25 scan points) areas during road surveys conducted in
Grand Teton National Park between June-October 2007-2010.
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
15
On average we saw < 1 cyclist passing on the road in the control per survey, peaking in 2008
(Figure 4). In the treatment, from 2007 to 2008, we observed an increase from 1 to 2 bicyclists
passing on the road on average per survey, then decreasing to <1 per survey in 2009 and 2010
coinciding with the opening of the pathway. With the opening of the pathway to public use, we
saw 8-9 bicyclists passing on the pathway per survey in 2009 and ca. 7 in 2010 (Figure 5), and
recorded 1-2 bicyclists and 2-3 pedestrians stopped on the pathway during each survey.
Figure 4. Mean number (± standard error) of bicylists on Teton Park Road within 200 meters of road survey
scan points in the control (n = 17scan points) and treatment (n=25 scan points) areas during road surveys
conducted in Grand Teton National Park between June-October 2007-2010.
Figure 5. Mean number (± standard error) of pathway users observed within 200 meters of treatment scan
points (n=25) during road surveys conducted in Grand Teton National Park between June-October 2009 and
2010.
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
16
We recorded at least one person watching wildlife, on average, at about 34% of the 17 control
scan points per survey in 2008, but wildlife viewing in the control decreased to ca. 5% and 3.5%
in 2009 and 2010, respectively (Figure 6). In the treatment, at least one person per survey was
observed viewing wildlife from the road at about 56% of the 25 treatment scan points in 2008,
decreasing to about 26% in 2009, and rebounding to 50% in 2010. People were seen watching
wildlife from the pathway as well; in 2008, prior to the completion of the pathway, we observed
people watching wildlife from the pathway at ca. 3% of the treatment scan points, and this
activity increased to ca. 7% and 12% in 2009 and 2010, respectively, after the pathway opened
for public use.
Figure 6. Mean percentage (± standard error) of road survey scan points in the control (n = 17 scan points)
and treatment (n=25 scan points) where at least one person was seen viewing wildlife during road surveys
conducted in Grand Teton National Park between June-October in 2008-2010.
Ungulate activities
During 421 road surveys, we recorded data on 2,319 ungulate groups, yielding 14,769 individual
ungulate observations, with 63% (n= 9,249) observed during day and 37% (n=5,520) during
crepuscular surveys. Elk were most commonly observed, comprising 57% of group observations
(1,304 elk groups) and 83% of all ungulate observations (12,134 elk observations). Pronghorn
comprised 32% of groups observed (728 pronghorn groups) and 14% of all ungulate
observations (2,039 pronghorn observations). Mule deer and moose observations comprised
only 6% (138 mule deer groups) and 5% (126 moose groups) of group observations, and 2%
(257 mule deer observed) and 1% (201 moose observed) of all ungulate observations. We
recorded 151 hours of ungulate behaviors over 670 focal animal samples collected in day (n=413
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
17
or 56% of focal samples) and crepuscular (n=319 or 43% of focal samples) periods before,
during and after road surveys.
Mule Deer – We recorded 257 mule deer observations within 138 mule deer groups (Table 2;
Appendix Map 1) during 104 road surveys (25% of 421 survey conducted). Group sizes ranged
from 1-9 individuals (mean±SE=1.86±0.12). Only 9% of all mule deer observed (19 groups
containing 24 mule deer) were recorded in the control area while most mule deer (233
individuals in 121 groups) were seen in the treatment area (Table 2).
Table 2. Summary of mule deer groups and number within groups observed at crepuscular and day periods
in the control and treatment areas during 421 road surveys conducted in Grand Teton National Park between
June-October 2007-2010.
YEAR
MULE DEER
CONTROL AREA
TREATMENT AREA
TO
TA
L
GR
OU
PS
TO
TA
L #
SE
EN
Crepuscular Day
Crepuscular Day
groups # seen groups # seen groups # seen groups # seen
2007 1 1 4 5
5 12 12 45
22 63
2008 1 3 4 8
8 18 17 28
30 57
2009 0 0 3 3
13 25 21 38
37 66
2010 1 1 3 3
11 18 34 49
49 71
2007-2010 3 5 14 19 37 73 84 160 138 257
Mule deer were seen more often per survey in the treatment in 2007, prior to pathway
construction, compared to other years (Figure 7, panel A). The oscillating trend in mule deer
group distance from the road in the control (Figure 7, panel B) is based on 19 group sightings
over the four years of the study, with annual means that ranged from ca. 84 meters to the TPR in
2009 (n=3) to 239 in 2008 (n=5). In the treatment, annual mean distance to the road ranged from
ca. 148 meters in 2008 (n=25) to 222 in 2009 (n=34). Data on behavioral responses in mule deer
were scarce and highly variable (Figure 7, panels C & D). The mean proportion of individuals
responding in the control was based on only 3, 1, and 3 group observations in 2007, 2008 and
2009, respectively; in 2010, we recorded no mule deer responding in 21 group behavior scan
observations in the treatment (Figure 7, panel C). Likewise, in the control, we were able to
obtain only 4 focal samples during the four years of the study, totaling only 51 minutes of
observation time. In the treatment, based on 8 hours of observations collected during 35 focal
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
18
samples, the proportion of time spent responding by mule deer in the treatment ranged from
approximately 16% to 32%, peaking in 2010 after the lowest level of observed responsiveness in
2008 (Figure 7, panel D).
Figure 7. Annual mean (± standard error) A) number mule deer viewed (standardized by viewable area) per
survey, B) distance of mule deer from road per group, C) proportion of mule deer responding per group, and
D) proportion time of mule deer responding per focal sample in the control and treatment areas in Grand
Teton National Park from June-October of 2007-2010.
Moose –Moose observations were also sparse. All but three moose groups observed were seen
in the treatment; in 2010, no moose were seen in the control during road surveys (Table 3;
Appendix Map 2). Moose group sizes ranged from 1 to 5 individuals (mean±SE=1.59±0.06).
Numbers of moose viewed per survey in the treatment appear stable in the first three years and
increased in 2010 (Figure 8, panel A). Moose were farther from the road in the treatment in
2008 (337±95.7m) compared to 2007 (145±53.8m), 2009 (186±33.6m) and 2010 (151±15.7m;
Figure 8, panel B). Behavioral responsiveness was highly variable in the treatment in 2007
(n=11) and 2009 (n=29); of 8 and 38 group behavior scan in 2008 and 2010, respectively, no
moose were responsive in the treatment (Figure 8, panel C). In the control, the 3 groups seen in
(A) (B)
(C) (D)
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
19
the first three years of the study moved out of view before we were able to conduct a behavior
scan. The proportion of time moose were observed responding during 46 focal samples totaling
over 9 hours of behavioral observations in the treatment ranged between approximately 12% in
2007 and 22% in 2010 (Figure 8, panel D).
Table 3. Summary of moose groups and number within groups observed at crepuscular and day periods in
the control and treatment areas during 421 road surveys conducted in Grand Teton National Park between
June-October 2007-2010.
YEAR
MOOSE
CONTROL AREA
TREATMENT AREA
TO
TA
L
GR
OU
PS
TO
TA
L #
SE
EN
Crepuscular Day
Crepuscular Day
groups # seen groups # seen groups # seen groups # seen
2007 0 0 1 1
3 5 10 17
14 23
2008 0 0 1 2
8 12 12 20
21 34
2009 0 0 1 1
15 23 24 37
40 61
2010 0 0 0 0
15 27 36 56
51 83
2007-2010 0 0 3 4 41 67 82 130 126 201
Figure 8. Annual mean (± standard error) A) number moose viewed (standardized by viewable area) per
survey, B) distance of moose from road per group, C) proportion of moose responding per group, and D)
proportion time moose responded per focal sample in the control and treatment areas in Grand Teton
National Park from June-October of 2007-2010.
(A) (B)
(C)
(D)
20
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
Elk – We viewed 1,321 elk groups comprised of 12,256 individual elk observations in the
treatment and control over the four years of the study (Table 4; Appendix Map 3). Elk groups
seen from the road ranged between 1 to 128 individuals (mean±SE = 9.28±0.41). Results of
quantitative analyses reviewed in the text below are limited to significant main effects and
year*area interactions; all results are included in referenced tables.
Table 4. Summary of elk groups and number within groups observed at crepuscular and day periods in the
control and treatment areas during 421 road surveys conducted in Grand Teton National Park between June-
October 2007-2010.
YEAR
ELK
CONTROL AREA
TREATMENT AREA
TO
TA
L
GR
OU
PS
TO
TA
L #
SE
EN
Crepuscular Day
Crepuscular Day
groups # seen groups # seen groups # seen groups # seen
2007 42 310 21 226
43 577 61 931
167 2084
2008 53 324 101 964
94 914 119 1255
367 3457
2009 54 319 90 536
139 1485 212 1756
495 4096
2010 44 158 32 232
7 828 139 1401
292 2619
2007-2010 193 1111 244 1958 283 3804 531 5343 1321 12256
Number of elk viewed — In early season, we observed more elk per survey during crepuscular
periods (mean±SE=0.0199±0.0034/ha) compared to day periods (0.0122±0.0012/ha; Figure 9;
see Table 5 for model outputs). Additionally, more elk were viewed in the control
(0.0202±0.0028/ha) than in the treatment (0.0106±0.0011/ha) during early season (Table 5).
Annual trends of number of elk viewed during early season also fluctuated significantly between
years throughout the study area, increasing from 2007 (0.0158±0.0075/ha) to 2008
(0.0217±0.0042/ha) and decreasing in 2009 (0.0144±0.0017/ha) and 2010 (0.0099±0.0016/ha;
Table 5). Annual trends in number of elk viewed also varied between the control and treatment
as reflected by the significant year*area interaction in early season (Table 5); the number of elk
viewed in the control significantly decreased between 2008 (0.0312±0.0072/ha) and 2010
(0.0088±0.0022/ha) while remaining relatively stable or increasing slightly in the treatment
during the study (Table 6, Figure 9). In late season, the number of elk viewed in the study area
per survey decreased annually (Table 5) from 2007 (0.0247±0.0062/ha) to 2008
(0.0204±0.0028), 2009 (0.0155±0.0030/ha) and 2010 (0.0121±0.0028/ha; Figure 9).
21
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
Figure 9. Annual mean (± standard error) number of elk viewed (standardized by viewable area) by season
and time of day in the control and treatment areas per road survey, and model-derived significant effects by
season, in Grand Teton National Park during June-October of 2007-2010.
CREPUSCULAR DAY
22
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
Table 5. Main and interactive effects influencing mean number of elk viewed/viewable per road survey
(n=421) over three seasons. Bold text highlights statistically significant (p<0.05) results (see Table 6 for post
hoc analyses of significant interaction). Road surveys were conducted in Grand Teton National Park between
June-October 2007-2010.
SEASON EFFECT Between
survey DF Within
survey DF F-Value P-value
Earl
y (
Ju
ne
-Ju
ly1
5)
YEAR 3 93 3.692 0.015
AREA 1 97 14.819 0.000
YEAR*AREA 3 93 4.152 0.008
TIME 1 97 5.280 0.024
YEAR*TIME 3 93 0.283 0.837
AREA*TIME 1 97 1.500 0.224
YEAR*AREA*TIME 3 93 0.971 0.410
Mid
(Ju
ly 1
6-A
ug
31)
YEAR 3 86 2.215 0.092
AREA 1 86 0.038 0.845
YEAR*AREA 3 86 1.535 0.211
TIME 1 86 0.007 0.932
YEAR*TIME 3 86 2.600 0.057
AREA*TIME 1 86 4.763 0.032
YEAR*AREA*TIME 3 86 1.749 0.163
Late
(S
ep
t-O
ct
15) YEAR 3 84 2.959 0.037
AREA 1 65 0.840 0.363
YEAR*AREA 3 65 1.362 0.262
TIME 1 85 3.110 0.081
YEAR*TIME 3 84 1.163 0.329
AREA*TIME 1 65 5.977 0.017
YEAR*AREA*TIME 3 65 2.343 0.081
Table 6. Post hoc contrasts exploring the significant early season year*area interaction in Table 5. Bold text
highlights statistically significant (p<0.05) results.
SEASON AREA
YEARS CONTRASTED
CONTRAST ESTIMATE
(A-B) SE DF T-value
Tukey HSD P-value A B
Earl
y (
Ju
ne
-Ju
ly1
5)
CO
NT
RO
L 2007 2008 -0.028 0.029 158 -0.950 0.778
2007 2009 0.008 0.028 158 0.280 0.992
2007 2010 0.053 0.029 158 1.852 0.256
2008 2009 0.036 0.018 158 1.937 0.220
2008 2010 0.081 0.020 158 4.158 0.000
2009 2010 0.046 0.018 158 2.558 0.058
TR
EA
TM
EN
T 2007 2008 -0.041 0.024 158 -1.697 0.331
2007 2009 -0.050 0.022 158 -2.277 0.111
2007 2010 -0.041 0.023 158 -1.796 0.282
2008 2009 -0.010 0.017 158 -0.566 0.942
2008 2010 0.000 0.018 158 -0.007 1.000
2009 2010 0.010 0.015 158 0.624 0.924
23
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
Elk distance from road — We measured distance from road for 1,304 elk groups seen during 421
road surveys (Figure 10). Time of day influenced the distance at which we observed elk groups
from the road over all seasons (Table 7). Elk were seen 218m, 329 m and 470m farther from the
road during day surveys compared to crepuscular surveys in early, mid and late seasons,
respectively. During early season, elk distances from the road fluctuated significantly over the
years of the study (Table 7), increasing from 2007 (mean±SE=341.3±48.7) to 2008
(580.6±32.2m) and 2009 (783.4±31.7m), and decreasing in 2010 (560.5±33.4m). During mid
season, elk groups were seen progressively farther from the road each year of the study (Table
7), from 2007 (438.9±81.8m) to 2008 (558.6±52.4m), 2009 (620.2±58.26m), and 2010
(694.1±60.4m). During late season, elk were, on average, seen about 418m farther from the road
in the treatment (745.4±37.3m) compared to the control (327.7±30.8m). Despite these effects,
no significant year*area interactions emerged (Table 7); annual trends were similar in the
treatment and control areas.
Table 7. Main and interactive effects influencing mean elk group distance from the road over three seasons
Bold text highlights statistically significant (p<0.05) results. Group distance data were recorded during 421
road surveys conducted in Grand Teton National Park between June-October 2007-2010.
SEASON EFFECT Between
groups DF Within
groups DF F-
Value P-value
Ea
rly
(J
un
e-J
uly
15
) Year 3 173 8.440 0.000
Area 1 217 0.420 0.518
Year*Area 3 173 1.771 0.154
Time 1 364 9.927 0.002
Year*Time 3 287 2.589 0.053
Area*Time 1 364 0.011 0.915
Year*Area*Time 3 287 1.323 0.267
Mid
(Ju
ly 1
6-A
ug
31
) Year 3 151 4.766 0.003
Area 1 163 2.611 0.108
Year*Area 3 151 2.114 0.101
Time 1 196 9.245 0.003
Year*Time 3 193 0.955 0.415
Area*Time 1 196 0.697 0.405
Year*Area*Time 3 193 0.435 0.728
La
te (
Se
pt-
Oc
t 1
5)
Year 3 144 0.647 0.586
Area 1 155 25.899 0.000
Year*Area 3 144 0.410 0.746
Time 1 205 26.070 0.000
Year*Time 3 202 0.604 0.613
Area*Time 1 205 1.236 0.267
Year*Area*Time 3 202 0.292 0.831
24
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
Figure 10. Annual mean (± standard error) elk group distance (meters) from the road by season and time of
day in the control and treatment areas, with model-derived significant effects by season, in Grand Teton
National Park during June-October of 2007-2010.
CREPUSCULAR DAY
25
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
Elk group scan responsiveness – During 421 road surveys, we recorded behavior scans on 419
elk groups, accounting for 3,722 elk behavior observations (Figure 11). In early season, we
detected no significant main or interactive effects (Table 8). In mid seasons, we conducted
behavior scans on 2, 3, and 8 elk groups in the control area in 2008, 2009 and 2010, respectively,
with no elk observed responding (Figure 11). Similarly, in late seasons, we recorded scan
samples where no elk responded in 6 and 8 groups in the control in 2009 and 2010, respectively
(Figure 11). Low sample sizes, however, precluded statistical analyses for these seasons.
Table 8. Main and interactive effects influencing probability of individual elk responding per group during
early seasons. Elk group behavior scan data were recorded during 421 road surveys conducted in Grand
Teton National Park between June-October 2007-2010.
SEASON EFFECT Between groups
DF
Within groups
DF
F-Value
P-value
Early (June-July15)
YEAR 3 28 0.701 0.559
AREA 1 15 1.466 0.244
YEAR*AREA 3 15 1.850 0.182
TIME 1 55 1.946 0.169
YEAR*TIME 3 56 0.343 0.794
AREA*TIME 1 105 1.413 0.237
YEAR*AREA*TIME 3 103 0.438 0.726
26
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
Figure 11. Annual mean (± standard error) proportion of elk per group observed behaviorally responding by
season and time of day in the control and treatment area, and model-derived significant effects by season, in
Grand Teton National Park during June-October of 2007-2010.
CREPUSCULAR DAY
27
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
Elk focal sample responsiveness – We conducted 330 elk focal samples, recording 78 hours of
focal elk behavior data over the four years of the study (Figure 12). In early season, the
probability of an elk responding was higher during crepuscular periods (mean±SE=0.231±0.037)
than day periods (0.165±0.030; Table 9). The three-way interaction year*area*time was
significant (Table 9); post hoc tests analyzing the effect of year, area and year*area by
crepuscular and day observation times separately, yielded no further significant results (Table
10), indicating that time of day was the driver of the significant three-way interaction. In mid
season, annual trends in responsiveness varied between control and treatment as represented by
the significant year*area interaction (Table 9). However, post hoc annual contrasts failed to find
significant differences between years (Table 11), likely due to high variability in these estimates
and conservative adjusted p-values associated with the Tukey’s multiple comparison test. Effect
plots suggest that behavioral responsiveness in elk in the treatment decreased while increasing in
the control, particularly during mid season crepuscular periods (Figure 12).
28
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
Figure 12. Annual mean (± standard error) percent time elk responded per focal sample, by season and time
of day in the control and treatment areas, with model-derived significant effects by season, in Grand Teton
National Park from June-October of 2007-2010.
CREPUSCULAR DAY
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
29
Table 9. Main and interactive effects influencing probability of an elk responding per second during 330 elk
focal sample collected in the control and treatment areas over three seasons. Bold text highlights statistically
significant (p<0.05) results (see Tables 10, 11 for post hoc analyses of significant interactions). Focal samples
were collected in Grand Teton National Park between June-October 2007-2010.
SEASON EFFECT Between
sample DF Within
sample DF F-Value P-value
Early (June-July15)
YEAR 3 24 0.376 0.771
AREA 1 27 0.013 0.910 YEAR*AREA 2 27 0.012 0.988
TIME 1 42 16.500 0.000
YEAR*TIME 2 43 2.447 0.099 AREA*TIME 1 40 0.085 0.772
YEAR*AREA*TIME 2 41 7.850 0.001
Mid (July 16-Aug 31)
YEAR 3 41 0.086 0.967
AREA 1 41 0.320 0.575
YEAR*AREA 3 27 4.012 0.018
TIME 1 44 1.000 0.323
YEAR*TIME 3 45 4.106 0.012
AREA*TIME 1 45 0.209 0.650
YEAR*AREA*TIME 1 45 0.507 0.480
Late (Sept-
Oct 15)
YEAR 3 68 0.692 0.560
AREA 1 81 0.089 0.767
YEAR*AREA 3 67 1.453 0.235
TIME 1 92 0.001 0.981
YEAR*TIME 3 169 8.851 0.000
AREA*TIME 1 92 0.123 0.727
YEAR*AREA*TIME 3 171 1.659 0.178
Table 10. Post hoc analysis of the significant early season three-way interaction (see Table 9) for the
probability of an elk responding per second of a focal sample, dropping time from the model to assess effects
at crepuscular and day periods separately.
EFFECT Between
sample DF Within
sample DF F-Value P-value
EARLY SEASON, CREPUSCULAR
YEAR 3 17 0.784 0.519
AREA 1 8 0.016 0.903
YEAR*AREA 2 8 0.026 0.974
EARLY SEASON, DAY
YEAR 3 22 0.236 0.870
AREA 1 18 0.227 0.639
YEAR*AREA 2 12 0.355 0.709
30
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
Table 11. Post hoc contrasts of annual year*area interactions (see Table 9) by area for midseason elk
probabilities of responding per second during focal samples.
AREA
YEARS CONTRASTED
CONTRAST ESTIMATE
(A-B) SE DF T-value
Tukey's HSD P-value
A B
CONTROL
2007 2008
Inestimable contrasts (high variability)
2007 2009
2007 2010
2008 2009
2008 2010
2009 2010 0.161 1.055 44 0.153 0.879
TREATMENT
2007 2008 -0.046 0.534 44 -0.087 1.000
2007 2009 1.206 0.583 46 2.070 0.189
2007 2010 0.446 0.834 44 0.535 0.950
2008 2009 1.252 0.591 45 2.119 0.173
2008 2010 0.493 0.840 43 0.586 0.935
2009 2010 -0.760 0.872 44 -0.871 0.820
Pronghorn antelope –Over the duration of the study, we saw 734 groups of pronghorn containing
2,055 individual pronghorn observations (Table 12; Appendix Map 4). Pronghorn groups ranged
in size from 1-22 individuals (mean±SE =2.79±0.1); most (>95%) pronghorn groups contained
<10 animals.
Table 12. Summary of pronghorn antelope groups and number within groups observed at crepuscular and
day periods in the control and treatment areas during 421 road surveys conducted in Grand Teton National
Park between June-October 2007-2010.
YEAR
PRONGHORN ANTELOPE
CONTROL AREA
TREATMENT AREA
TO
TA
L
GR
OU
PS
TO
TA
L #
SE
EN
Crepuscular Day
Crepuscular Day
groups #
seen groups
# seen
groups #
seen groups
# seen
2007 3 8 5 12
12 79 40 191
60 290
2008 12 40 48 127
17 87 71 271
148 525
2009 10 33 45 104
51 111 205 484
311 732
2010 13 38 66 112
21 64 115 294
215 508
2007-2010 38 119 164 355 101 341 431 1240 734 2055
31
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
Number of pronghorn viewed – During early season, the number of pronghorn viewed per survey
varied significantly over years (Figure 13, Table 13); we saw the fewest pronghorn per survey in
2007 (mean±SE=0.0025±0.0006/ha), the most in 2008 (0.0048±0.0005/ha), with moderate
numbers in 2009 (0.0037±0.0003/ha) and 2010 (0.0026±0.0003/ha). The number of pronghorn
viewed was significantly higher in the control (0.0043±0.0004/ha) than in the treatment area
(0.0030±0.0002/ha; Table 13). Post hoc tests exploring the significant early season three-way
interaction, analyzing crepuscular and day observations separately, did not reveal any significant
year*area interactions (Table 14). During early season crepuscular periods, we saw significantly
more pronghorn in the control (0.0052 ± 0.0008/ha) than in the treatment (0.0029±0.0004/ha).
During day surveys, the number of pronghorn viewed across both the control and treatment areas
together was lowest in 2007 (mean±SE= 0.0019±0.0004/ha), peaked in 2008 (0.0048±
0.0006/ha), and was intermediate in 2009 (0.0038±0.0004/ha) and 2010 (0.0023±0.0003/ha). In
late season, we saw the most pronghorn in 2007 (0.0043±0.0005/ha), dropping in 2008
(0.0037±0.0006/ha) and 2009 (0.0015±0.0004/ha), and increasing in 2010 (0.0025±0.0004/ha;
Table 13). There were no significant year*area interactions (Table 13), suggestive of no
pathway impact.
32
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
Figure 13. Annual mean (± standard error) number of pronghorn viewed (standardized by viewable area) by
season and time of day in the control and treatment areas per road survey, with model-derived significant
effects by season, in Grand Teton National Park during June-October of 2007-2010.
Year p=0.020
CREPUSCULAR DAY
33
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
Table 13. Main and interaction effects influencing numbers of pronghorn viewed (standardized by viewable
area) from the Teton Park Road in the control and treatment areas per road survey over three seasons. Bold
text highlights statistically significant (p<0.05) results (see Table 14 for post hoc analyses of significant
interaction). Road surveys were conducted in Grand Teton National Park between June-October 2007-2010.
SEASON EFFECT Between
survey DF Within
survey DF F-
Value P-value
Earl
y (
Ju
ne
-Ju
ly1
5) YEAR 3 121 3.912 0.010
AREA 1 110 6.325 0.013
YEAR*AREA 3 101 1.309 0.276
TIME 1 130 1.642 0.202
YEAR*TIME 3 121 1.906 0.132
AREA*TIME 1 110 3.276 0.073
YEAR*AREA*TIME 3 101 2.771 0.045
Mid
(Ju
ly 1
6-A
ug
31)
YEAR 3 88 1.367 0.258
AREA 1 89 0.849 0.359
YEAR*AREA 3 76 1.290 0.284
TIME 1 94 0.495 0.483
YEAR*TIME 3 88 0.182 0.909
AREA*TIME 1 86 0.169 0.682
YEAR*AREA*TIME 2 70 0.520 0.597
Late
(S
ep
t-O
ct
15) YEAR 3 71 3.483 0.020
AREA 1 74 2.358 0.129
YEAR*AREA 3 53 1.310 0.281
TIME 1 73 0.106 0.746
YEAR*TIME 3 71 1.426 0.242
AREA*TIME 1 47 0.002 0.960
YEAR*AREA*TIME 2 40 1.364 0.267
Table 14. Post hoc analysis of the significant early season three-way interaction (see Table 13) effect on
pronghorn viewed, dropping time from the model to assess effects at crepuscular and day periods separately.
EFFECT between
survey DF within
survey DF F-Value P-value
EARLY SEASON, CREPUSCULAR OBSERVATIONS
YEAR 3 29 0.743 0.535
AREA 1 29 7.000 0.013
YEAR*AREA 3 28 2.147 0.117
EARLY SEASON, DAY OBSERVSATIONS
YEAR 3 85 7.703 0.000
AREA 1 79 0.334 0.565
YEAR*AREA 3 67 0.288 0.834
34
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
Pronghorn distance from road —During early season, pronghorn were farther from the road in
the control (mean±SE=443.4±33.1m) than in the treatment (355.7±21.5m; Figure 14). In mid
season, pronghorn observations during the day were, on average, farther from the road
(308.6±22.7m) than groups observed during crepuscular periods (216.5±28.4m; Table 15).
Additionally, during mid season, annual trends in pronghorn distance from road in the treatment
and control diverged as indicated by the significant year*area interaction (Table 15). Post hoc
contrasts (Table 16) confirm that, during mid season, pronghorn were observed about 164m
farther from the road in the treatment in 2010 (366.9±50.2m) than 2007 (202.2±35.3m), while
pronghorn shifted, on average, about 181m closer to the road in the control in 2010
(408.9±69.7m) compared to 2009 (228.2±63.9m).
Table 15. Main and interactive effects influencing mean pronghorn distances from the Teton Park Road over
three seasons. Bold text highlights statistically significant (p<0.05) results (see Table 16 for post hoc analyses
of significant interaction). Pronghorn distances to road were collected during road surveys conducted in
Grand Teton National Park between June-October 2007-2010.
SEASON EFFECT Between
groups DF Within
groups DF F-Value P-value
Earl
y (
Ju
ne
-Ju
ly1
5) YEAR 3 295 0.498 0.684
AREA 1 359 4.123 0.043
YEAR*AREA 3 294 1.128 0.338
TIME 1 372 2.223 0.137
YEAR*TIME 3 303 1.212 0.306
AREA*TIME 1 256 0.053 0.818
YEAR*AREA*TIME 2 255 0.388 0.679
Mid
(Ju
ly 1
6-A
ug
31) YEAR 3 130 1.030 0.382
AREA 1 162 0.264 0.608
YEAR*AREA 3 161 5.137 0.002
TIME 1 145 6.330 0.013
YEAR*TIME 3 136 0.554 0.646
AREA*TIME 1 162 1.083 0.300
YEAR*AREA*TIME 2 160 0.412 0.663
Late
(S
ep
t-O
ct
15)
YEAR 3 83 0.031 0.993
AREA 1 101 2.082 0.152
YEAR*AREA 3 98 2.034 0.114
TIME 1 95 2.705 0.103
YEAR*TIME 3 93 0.604 0.614
AREA*TIME 1 98 0.042 0.838
YEAR*AREA*TIME 2 96 1.020 0.365
35
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
Table 16. Post hoc annual contrasts of the mid season year*area interaction effects (see Table 15) on
pronghorn group distance from road.
SEASON AREA YEARS CONTRAST
(A-B) SE DF T-value
Tukey HSD P-value A B
Mid
(Ju
ly 1
6-A
ug
31)
CO
NT
RO
L
2007 2008
Inestimable contrast due to low sample size: n=2 observations in the control in 2007 mid season during midday
2007 2009
2007 2010
2008 2009 -5.049 2.519 156 -2.005 0.114
2008 2010 1.664 2.683 159 0.620 0.809
2009 2010 6.713 2.623 161 2.560 0.030
TR
EA
TM
EN
T 2007 2008 -2.863 3.471 158 -0.825 0.843
2007 2009 -3.045 1.804 107 -1.688 0.333
2007 2010 -5.185 1.958 112 -2.649 0.044
2008 2009 -0.182 3.267 156 -0.056 1.000
2008 2010 -2.322 3.354 155 -0.692 0.900
2009 2010 -2.140 1.567 70 -1.365 0.523
Pronghorn group scan responsiveness – We conducted 412 behavior scans on pronghorn groups
during the four years of the study. High variability in pronghorn group scan observations (Figure
15) combined with low sample sizes in some years precluded statistical analyses in early and mid
season. No significant main or interactive effects were evident for the late season (Table 17).
Graphical examination also do not suggest any clear pathway impact (Figure 15).
Table 17. Main and interaction effects influencing percent of pronghorn behaviorally responding, per group
scan observation, during road surveys on Teton Park Road during late season. Road surveys were conducted
in Grand Teton National Park during June-October 2007-2010.
EFFECT: Late Season (Sept-Oct 15)
Between groups
DF
Within groups
DF F-Value P-value
YEAR 3 35 1.19535 0.3256
AREA 1 46 0.10097 0.7521
YEAR*AREA 2 44 0.10655 0.8992
TIME 1 43 0.23243 0.6322
YEAR*TIME 3 35 1.71596 0.1813
AREA*TIME 1 46 0.18185 0.6718
YEAR*AREA*TIME 2 44 0.03956 0.9612
36
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
Figure 14. Annual mean (± standard error) pronghorn antelope group distances (meters) to road by season
and time of day in the control and treatment areas, with model-derived significant effects by season, in Grand
Teton National Park during June-October of 2007-2010.
CREPUSCULAR DAY
37
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
Figure 15. Annual mean (± standard error) proportion of pronghorn observed behaviorally responding per
group by season and time of day in the control and treatment area in Grand Teton National Park during
June-October of 2007-2010.
Pronghorn focal sample responsiveness – We conducted 251 focal samples on pronghorn,
obtaining more than 53 hours of observation time during the study. Analyses for all seasons
revealed no significant effects (Table 18). Graphical examination of mean proportion of time
pronghorn responded per focal sample also does not suggest a clear pathway impact (Figure 16).
CREPUSCULAR DAY
38
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
Figure 16. Annual mean (± standard error) percent time pronghorn responded per focal sample, by season
and time of day in the control and treatment areas, with model-derived significant effects by season, in Grand
Teton National Park from June-October of 2007-2010.
CREPUSCULAR DAY
39
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
Table 18. Main and interaction effects influencing pronghorn percent time responding per focal behavior
sample over three seasons. Bold text highlights statistically significant (p<0.05) effects.
SEASON EFFECT Between sample
DF
Within sample
DF F-Value P-value
Early (June-July15)
YEAR 3 67 0.791 0.503
AREA 1 57 0.548 0.462
YEAR*AREA 2 56 2.339 0.106
TIME 1 73 1.831 0.180
YEAR*TIME 2 74 0.513 0.601
AREA*TIME 1 71 6.481 0.013
YEAR*AREA*TIME 2 72 0.783 0.461
Mid (July 16-Aug 31)
YEAR 3 56 0.738 0.534
AREA 1 56 0.914 0.343
YEAR*AREA 2 56 0.558 0.576
TIME 1 57 0.507 0.480
YEAR*TIME 3 59 2.509 0.068
AREA*TIME 1 56 0.984 0.325
YEAR*AREA*TIME 1 56 1.421 0.238
Late (Sept- Oct 15)
YEAR 3 36 0.152 0.928
AREA 1 28 2.216 0.148
YEAR*AREA 2 28 0.137 0.873
TIME 1 40 0.009 0.925
YEAR*TIME 3 40 3.496 0.024
AREA*TIME 1 36 0.229 0.635
YEAR*AREA*TIME 2 35 3.182 0.054
DISCUSSION
The new multi-use pathway along Teton Park Road (TPR) resulted in direct habitat loss,
construction disturbance and changes in human activities along the transportation corridor in
Grand Teton National Park. Our results, however, did not clearly or consistently demonstrate
that ungulate distribution and behavior, or wildlife viewing opportunities, were considerably
affected by the construction and use of the pathway. The strongest evidence of an ungulate
avoidance response to the pathway included a relatively modest shift in pronghorn distribution
away from the road and path corridor during periods of peak visitation and pathway use.
Additionally, elk behavioral responsiveness during peak visitation appeared to decrease in the
40
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
treatment relative to the control, particularly during crepuscular hours, potentially suggestive of
elk tolerance to activities in the treatment. Ultimately, park visitors could still view these species
from the TPR during pathway construction and after the pathway opened.
Although automobile traffic volumes did not change over the duration of the study (Sawyer et al.
2011), the installation of the pathway increased and diversified non-motorized travel activities in
the TPR corridor, meeting a stated objective of the park transportation plan (National Park
Service 2006). With the opening of the pathway, most pedestrian and bicyclist activities shifted
from the road to the pathway and we estimated a concurrent ~3 fold increase in bicycling
activity. Infrared trail counters detected as many as 148 pathway users passing on the pathway
per hour, peaking seasonally between June 15 and August 30, and daily between 1100 and 1600
hours (Costello et al. 2011). This shift in non-motorized travel from the road to the pathway
presumably decreased the potential for collisions and conflicts between automobiles and
pedestrians and bicyclists, meeting another stated objective of the transportation plan (National
Park Service 2006). However, the overall increase in non-motorized travelers since the pathway
opened could potentially increase interactions and possibly conflicts between wildlife and park
visitors. This may be particularly true considering the three-fold increase in off-road and off-
trail travel in the treatment since pathway construction. Previous research demonstrates
ungulates are more likely to respond to off-road and off-path activities than the more frequent
but predictable activities that occur on linear infrastructure (MacArthur et al. 1979, Cassirer et al.
1992, Knight and Cole 1995, Miller et al. 2001, Papouchis et al. 2001, Borkowski et al. 2006).
Wildlife viewing opportunities did not seem to be substantially impacted by the pathway. Even
prior to the completion of pathway construction in 2008, people ventured from the road to watch
wildlife from the pathway at 3% of our sampling sites, and we measured a four-fold increase in
wildlife viewing from the pathway once it opened to the public. We documented people
watching wildlife from the road at more than 50% of our sampling points in the treatment area in
2008, decreasing to a quarter of our sampling points in 2009, and recovering in 2010 to a level
similar to that seen in 2008. In the control, by comparison, we witnessed people viewing
wildlife from the road at 34% of survey scan points in 2008, when ungulate observations in the
control peaked, but wildlife viewing dropped to 5% and 3.5% of the control scan points in 2009
41
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
and 2010, coinciding with a decrease in ungulate observations in the control (discussed below).
The fact wildlife viewing in the treatment in 2010 was similar to 2008 indicates that the pathway
did not have a considerable impact on visitor opportunities to see ungulates from the road. A
concurrent study assessing elk habitat use and movements based on fine-scale GPS data from a
sample of individual elk occupying the pathway study area indicated that high-use elk habitats
visible from the TPR did not noticeably change before, during and two years after pathway
construction (Sawyer et al. 2011). This corroborates our finding that opportunities to view
ungulates (elk, in particular) from the TPR did not change noticeably with the onset of pathway
construction and use, and foreshadows our findings of few detectable changes in ungulate
distribution which may be attributed to pathway activities.
If ungulates were avoiding pathway activities, we predicted a decreasing trend in the number of
ungulates seen per survey in the treatment compared to the control after pathway construction.
Our results do not support this prediction. For elk, we did not find statistical evidence for a
decrease in elk numbers or displacement away from the road in the treatment compared to the
control during the study. The only statistically significant year*area interaction, indicative of a
pathway impact in our BACI experimental design, was a decrease in the number of elk viewed in
the control between 2008 and 2010 during early season, while the number in the treatment area
did not significantly change over the same period. This lack of a pathway effect is again
consistent with findings from the concurrent study of GPS-collared elk that concluded that elk
movement, habitat use and crossings of the TPR were not affected by pathway construction or
use (Sawyer et al. 2011). We also predicted that we would see increased behavioral
responsiveness in the treatment compared to the control if the pathway induced an avoidance
response in ungulates. Again, contrary to this prediction, during the mid season when visitation
peaked, elk behavioral responsiveness during crepuscular periods appeared to decrease in the
treatment relative to the control over the duration of the study. Based on these outcomes, it was
not apparent that elk avoided the pathway or were substantially disturbed by human activities
occurring in the treatment during and after the construction of the pathway. Rather, it appears
elk may have been tolerant of human activities, including pathway construction and use in the
treatment.
42
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
Our results suggest that pronghorn may have avoided pathway activities by shifting away from
the TPR in the treatment. Specifically, in mid season, during peak visitation, pronghorn were
seen about 164 meters farther from the road in the treatment in 2010 compared to 2007, although
they were seen 97 meters closer, on average, to the road in the control region in 2010 compared
to 2009. This suggests that pronghorn may have been somewhat disturbed by construction and
pathway activities, occupying areas farther from the road corridor after pathway construction.
The estimated number of pronghorn viewed and measures of pronghorn behavioral
responsiveness, however, were variable and not indicative of a pathway effect, although it is
possible our methods failed to distinguish subtle pronghorn responses to pathway activities.
Compared to elk, pronghorn are considered to be more sensitive to human disturbances,
exhibiting risk-avoidance behavior in proximity to roads with traffic (Berger et al. 1983, Gavin
and Komers 2006). A study of the effect of anthropogenic noise on elk and pronghorn in the
same GTNP pathway study area in 2008 affirmed our findings of a higher level of behavioral
responsiveness in pronghorn compared to elk (Brown et al. In review.). Despite the relative
differences between elk and pronghorn responsiveness, both species were less likely to
behaviorally respond as noise levels and vehicle traffic intensified, which, as reviewed below,
potentially suggests habituation to such disturbances (Brown et al. In review.).
Compared to pronghorn and elk, far fewer mule deer were seen in the study area, comprising
only 2% of all ungulate sightings. We saw mule deer during a quarter of our road surveys, and
all but 9% of these observations were collected in the treatment area. Mean distances of mule
deer groups to the road in the treatment ranged from 148m to 222 m during 2008 and 2009,
respectively, with intermediate distances observed at the beginning and end of the study, offering
no obvious pattern of avoidance of pathway construction or use. Behavioral responsiveness of
mule deer was relatively infrequent and also not suggestive of a pathway impact. Mule deer
often bedded within sight of the road along edges of forested cover, well-camouflaged and
undetectable to most people on the road, as we observed many people pass by seemingly
unaware of the mule deer viewing opportunity. Despite their infrequent appearance near the
TPR, people were seen watching mule deer from the road during all four years of the study. Low
sample sizes precluded statistical analyses of mule deer observations, but the fact that we
continued to see mule deer in the treatment area during and after construction, with relatively
43
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
little behavioral responsiveness to human disturbance, suggests that this species did not appear to
be avoiding pathway activities.
Moose were observed the least often, comprising only 1% of all ungulate observations. Out of
421 road surveys, we saw only 4 moose in the control in the first three years of the study; no
moose were seen in the control in 2010, when annual moose sightings peaked in the treatment
(83 moose in 51 group observations; 41% of all moose observations). Moose were observed
farther from the road in the treatment in 2008 compared to other years, with mean distances to
the road decreasing to pre-pathway measurements by 2010. Although speculative, this may
indicate a response to pathway construction, followed by increasing tolerance to human activities
after the initial construction disturbance subsided and pathway use ensued. As with mule deer,
behavioral responsiveness of moose was variable and not indicative of a pathway effect.
Notably, moose prompted numerous and large gatherings of wildlife watchers, particularly
where the TPR crossed the Snake River in Moose, Wyoming. A large number of our off-road
and off-pathway observations were attributed to visitors venturing down along the river to see
moose in this area. Even with off-road and off-path activities and direct approaches by people,
moose opted to stay, sometimes for weeks, near the TPR in the treatment, offering visitors moose
viewing opportunities from the road. Thus, moose did not appear to be notably disturbed or
displaced by human activities in the treatment, including pathway activities in particular.
Wildlife responses to human activities are complex, but previous history of an animal’s exposure
to human activity plays an important role in understanding current responses to potential
disturbances (Bejder et al. 2006, Stankowich 2008, Bejder et al. 2009). In park settings, or in
areas where hunting is prohibited and people do not approach wildlife directly, animals that are
exposed to high-use human activity areas such as park roads have shown lower levels of
responsiveness to human activities than might be expected (Thompson and Henderson 1998,
Burson et al. 2000, Papouchis et al. 2001, Borkowski et al. 2006). The Teton Park Road has
afforded access and wildlife viewing opportunities to millions of park visitors for decades. Prior
to pathway construction in 2008, ungulates were visible from the road, and many visitors stopped
on the road to view elk and pronghorn, in particular. Assuming a majority of the ungulates
observed in the study area were habitual summer residents that annually returned to these
44
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
habitats, as was demonstrated for elk via GPS-collar movement data during the study (Sawyer et
al. 2011), these animals would have been familiar with relatively predictable and common
human activities that occur in the park’s transportation corridor. Frequent exposure to
predictable activities that result in neutral outcomes can induce habituation, a waning of response
to inconsequential stimulus (Eibl-Eibesfeldt 1970). This learned response is a behavioral
adaptation that allows animals to dedicate attention and energy toward fitness-enhancing
behaviors such as feeding, grooming, resting and mating rather than expending energy to flee
non-threatening activities (Thompson and Henderson 1998). Ungulates have been known to
habituate to regular exposure of non-lethal human activities (Stankowich 2008). Elk in particular
have shown habituation patterns along roads and other areas disturbed by human activities (Lyon
and Ward 1982, Morrison et al. 1995, Thompson and Henderson 1998).
Ungulates that occupied habitat visible from park roads may not be representative of the entire
population of ungulates in the larger region. There is potentially a self-segregating contingent of
ungulates from the larger population that were less tolerant of human activities and actively
avoided habitats seen from park roads even prior to the onset of this study (Bejder et al. 2006,
Vistnes and Nellemann 2007). Nevertheless, for ungulates that do not avoid roads, such as those
observed in our study, lack of displacement from human disturbance and reduced behavioral
responsiveness of individual animals can have negative impacts on fitness and population
persistence. For example, animals that do not exhibit strong behavioral avoidance of humans
may still suffer fitness impacts if the human disturbance is substantial but the costs of moving to
avoid it are overly high (George and Crooks 2006). Decreased behavioral responsiveness in
ungulates may also reduce their ability to visually detect predators and other potential threats
(Brown et al. In review.). Moreover, although tolerance to human activities may provide
opportunities for park visitors to view ungulates from the road, it may also lead to increased
human conflict such as negative encounters with recreationists (Olliff and Caslick 2003) or
collisions with vehicles (Ament et al. 2008), major concerns for park managers. Long-term
studies would be necessary to evaluate the actual demographic impacts of pathway activity on
ungulate populations, within and beyond habitats that can be seen from the road.
45
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
ACKNOWLEDGEMENTS
The National Park Service (NPS) funded this study through a grant from the Federal Highway
Administration, administered by the Cooperative Ecosystems Study Unit. Additional funding
support was provided by Rocky Mountain Goat Foundation, Anheuser-Busch Environmental
Fellowship, The Hill Memorial Fellowship, and Oscar and Isabel Anderson Graduate
Scholarship. We sincerely appreciate guidance and support provided by NPS personnel Steve
Cain, Sarah Dewey, Cindy O’Neill, John Stephenson, Margaret Wilson, and Sue Wolff.
Colorado State University (CSU) staff Carl Davis, James Frantz, Joyce Pratt and Val Romero
provided administrative support. Special thanks to Phil Chapman at CSU for statistical
consultation; Cara Ostrum, Matthew Imig, Libo Sun and Amy Davis provided assistance with
statistical analyses. Andre Breton provided valuable database assistance and advice. Academic
committee members Lisa Angeloni, Peter Newman, Tara Teel, and Dave Theobald provided
valuable feedback and advice on study design. Numerous field assistants helped collect field
observations: Rebecca Blaskovich, David Ellsworth, Daniel Kinka, Kathleen Knighton, Kendra
Martinez, Tim Sullivan and Kristin Van Ort. Lindsay Simpson assisted with organizing and
summarizing literature. We especially appreciate Casey Brown’s field support and collaboration
in 2008. Melinda Scott deserves special recognition for three years of field support, input on
study design and protocols, streamlining data collection and management and refinement of the
viewshed analyses. Jesse Barber, Chris Burdett, Steve Cain, Sarah Dewey, and Roger Dodds
provided helpful comments on an earlier version of this report.
LITERATURE CITED
Ament, R., A. Clevenger, O. Yu, and A. Hardy. 2008. An Assessment of Road Impacts on Wildlife Populations in
U.S. National Parks. Environmental Management 42:480-496.
Barber, J. R., K. R. Crooks, and K. M. Fristrup. 2010. The costs of chronic noise exposure for terrestrial organisms.
Trends in Ecology & Evolution 25:180-189.
Bejder, L., A. Samuels, H. Whitehead, H. Finn, and S. Allen. 2009. Impact assessment research: use and misuse of
habituation, sensitisation and tolerance in describing wildlife responses to anthropogenic stimuli. Marine
Ecology Progress Series 395:177-185.
Bejder, L., A. Samuels, H. Whitehead, and N. Gales. 2006. Interpreting short-term behavioural responses to
disturbance within a longitudinal perspective. Animal Behaviour 72:1149-1158.
Berger, J., D. Daneke, J. Johnson, and S. H. Berwick. 1983. Pronghorn foraging economy and predator avoidance in
a desert ecosystem: Implications for the convertion of large mammalian herbivores. Biological
Conservation 25:193-208.
Borkowski, J. J., P. J. White, R. A. Garrott, T. Davis, A. R. Hardy, and D. J. Reinhart. 2006. Behavioral responses of
bison and elk in Yellowstone to snowmobiles and snow coaches. Ecological Applications 16:1911-1925.
46
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
Brown, C. L., A. R. Hardy, J. R. Barber, K. M. Fristrup, K. R. Crooks, and L. M. Angeloni. In review. The effect of
human activities and their associated noise on ungulate behavior. Animal Behaviour.
Burson, S. L., III, J. L. Belant, K. A. Fortier, and W. C. Tomkiewicz, III. 2000. The effect of vehicle traffic on
wildlife in Denali National Park. Arctic 53:146.
Cassirer, E. F., D. J. Freddy, and E. D. Ables. 1992. Elk Responses to Disturbance by Cross-Country Skiers in
Yellowstone National Park. Wildlife Society Bulletin 20:375-381.
Childress, M. J. and M. A. Lung. 2003. Predation risk, gender and the group size effect: does elk vigilance depend
upon the behaviour of conspecifics? Animal Behaviour 66:389-398.
Costello, C. M., S. L. Cain, R. M. Nielson, C. Servheen, and C. C. Schwartz. 2011. Impacts of a multi-use pathway
on American black bears in Grand Teton National Park, Wyoming. Report for National Park Service,
Grand Teton National Park, Wyoming.
Eibl-Eibesfeldt, I. 1970. Ethology: the biology of behavior. Holt, Rinehart & Winston, New York, New York.
Evink, G. L. 2002. Interaction between roadways and wildlife ecology. Transportation Research Board, Washington,
D.C.
Forman, R. T. T., D. Sperling, J.A. Bissonette, A.P. Clevenger, C.D Cutshall, V.H. Dale, L. Fahrig, R. France, C.R.
Goldman, K. Heanue, J.A. Jones, F.J. Swanson, T. Turrentine, T.C. Winter. 2003. Road ecology science
and solutions. Island Press, Washington, DC.
Gavin, S. D. and P. E. Komers. 2006. Do pronghorn (Antilocapra americana) perceive roads as a predation risk?
Canadian Journal of Zoology 84:1775-1780.
George, S. L. and K. R. Crooks. 2006. Recreation and large mammal activity in an urban nature reserve. Biological
Conservation 133:107-117.
Green, R. H. 1993. Application of repeated measures designs in environmental impact and monitoring studies.
Austrailian Journal of Ecology 18:81-98.
Knight, R. L. and D. N. Cole. 1995. Factors that influence wildlife responses to recreationists. Pages 70-79 in R. L.
Knight and K. J. Gutzwiller, editors. Wildlife and recreationists: coexistence through management and
research. Island Press, Washington DC.
Lyon, L. J. and A. L. Ward. 1982. Elk and land management. Pages 442-477 in J. W. Thomas and D. E. Toweill,
editors. Elk of North America: ecology and management. Stackpole Books, Pennsylvania.
MacArthur, R. A., R. H. Johnston, and V. Geist. 1979. Factors influencing heart rate in free-ranging bighorn sheep:
a physiological approach to the study of wildlife harassment. Canadian Journal of Zoology 57:2010-2021.
Miller, S. G., R. L. Knight, and K. M. Clinton. 2001. Wildlife Responses to Pedestrians and Dogs. Wildlife Society
Bulletin 29:124-132.
Morrison, J. R., W. J. d. Vergie, A. W. Alldredge, A. E. Byrne, and W. W. Andree. 1995. The Effects of Ski Area
Expansion on Elk. Wildlife Society Bulletin 23:481-489.
National Park Service. 2006. Transportation Plan Final Environmental Impact Statement. Page 368, Moose,
Wyoming.
Olliff, T. and J. Caslick. 2003. Wildlife-human conflicts in Yellowstone: when animals and people get too close.
Pages 18-22 Yellowstone Science. National Park Service, Yellowstone National Park.
Papouchis, C. M., F. J. Singer, and W. B. Sloan. 2001. Responses of Desert Bighorn Sheep to Increased Human
Recreation. The Journal of Wildlife Management 65:573-582.
Riley, S. P. D., J. P. Pollinger, R. M. Sauvajot, E. C. York, C. Bromley, T. K. Fuller, and R. K. Wayne. 2006. A
southern California freeway is a physical and social barrier to gene flow in carnivores. Molecular Ecology
15:1733-1741.
Roedenbeck, I. A., L. Fahrig, C. S. Findlay, J. E. Houlahan, J. A. G. Jaeger, N. Klar, S. Kramer-Schadt, and E. A. v.
d. Grift. 2007. The Rauischholzhausen agenda for road ecology. Page 11 Ecology and Society.
Sawyer, H., R. M. Nielson, F. Hornsby, and L. McManus. 2011. Grand Teton National Park pathway elk study.
Western Ecosystems Technology, Inc., Laramie, Wyoming.
Smith, E. P. 2002. BACI design. Pages 141-148 in A. H. El-Shaarawi and W. W. Piegorsch, editors. Encyclopedia
of Environmetrics. John Wiley & Sons, Ltd., Chichester.
Stankowich, T. 2008. Ungulate flight responses to human disturbance: A review and meta-analysis. Biological
Conservation 141:2159-2173.
Stewart-Oaten, A. and J. R. Bence. 2001. Temporal and spatial variation in environmental impactg assessment.
Ecological Monographs 71:305-339.
Stewart-Oaten, A., W. W. Murdoch, and K. R. Parker. 1986. Environmental Impact Assessment:
"Pseudoreplication" in Time? Ecology 67:929-940.
47
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
Thompson, M. J. and R. E. Henderson. 1998. Elk Habituation as a Credibility Challenge for Wildlife Professionals.
Wildlife Society Bulletin 26:477-483.
Trombulak, S. C. and C. A. Frissell. 2000. Review of Ecological Effects of Roads on Terrestrial and Aquatic
Communities. Conservation Biology 14:18-30.
U.S. Naval Observatory. 2010. Sun or moon rise/set table. United States Naval Meteorology and Oceanography
Command, Washington DC.
Vistnes, I. and C. Nellemann. 2007. Impacts of human activity on reindeer and caribou: The matter of spatial and
temporal scales. Rangifer 12:47-56.
48
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
Appendix Map 1. Locations of mule deer groups observed from Teton Park Road between Moose and
Spalding Bay Road during road surveys conducted between June and October 2007-2010 in Grand Teton
National Park.
49
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
Appendix Map 2. Locations of moose groups observed from Teton Park Road between Moose and Spalding
Bay Road during road surveys conducted between June and October 2007-2010 in Grand Teton National
Park.
50
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
Appendix Map 3. Locations of elk groups observed from Teton Park Road between Moose and Spalding Bay
Road during road surveys conducted between June and October 2007-2010 in Grand Teton National Park.
51
Hardy and Crooks Grand Teton Pathways Final Report
September 15, 2011
Appendix Map 4. Locations of pronghorn antelope groups observed from Teton Park Road between Moose
and Spalding Bay Road during road surveys conducted between June and October 2007-2010 in Grand Teton
National Park.